The experiments proposed here are aimed at defining the neural substrates and functional properties of attentional selection mechanisms in the sensory pathways of the human brain. A major emphasis will be on identifying the levels of cortical processing at which relevant visual stimuli are preferentially selected from the environment. The principal method to be used will be multichannel recordings of event-related brain potentials (ERPs), combined in some experiments with functional magnetic resonance imaging (fMRI) and magneto encephalography (MEG) in order to improve anatomical specificity. Neural activity modulated by attention will be localized with respect to the retinotopic visual areas V1, V2, V3/VP, V3A, and V4 and to more anterior visual areas of the dorsal and ventral processing streams. Eight experiments are proposed that will investigate basic mechanisms of visual attention in tasks where stimuli are selected on the basis of location (spatial attention), ensembles of features (object-based attention), and cues in other modalities (cross-modal attention). These experiments aim to clarify a number of outstanding issues in attention research: (1) the role of primary visual cortex in spatial and object-based attention, (2) the possible role of EEG phase-locking in visual attention, (3) the neural mechanisms that enable the splitting of the "spotlight" of spatial attention, (4) the neural substrates of feature selection and integration during object-based attention, (5) the neural basis of attention-induced binocular rivalry, and (6) the neural interactions that underlie the cross-modal facilitation of perceptual processing. The long-range goal of this research is to improve mental health, particularly disturbances of selective attention that are characteristic of several clinical disorders such as schizophrenia, autism, and attention-deficit hyperactivity disorder. The studies proposed here will lead to an improved understanding of the basic mechanisms of both normal and disordered attention.